From pikuwww.mondal at gmail.com Wed Nov 6 15:01:16 2013 From: pikuwww.mondal at gmail.com (Wasim Mondal) Date: Wed, 6 Nov 2013 19:31:16 +0530 Subject: [Wannier] reliability for convergence Message-ID: Dear wannier experts I have run one calculation of cubic peroskite material to generate the Hamiltonian. After completing the calculation I am not able to understand my results are reliable or not. Please make some comments what I have done is correct or wrong.The detailed of the convergence is given below: user at shell:~/Material/cubic/final_cal2/subdir_2$ grep SPRD subdir_2.wout O_D= 42.9797537 O_OD= 1.9159766 O_TOT= 93.0933356 <-- SPRD O_D= 56.8597829 O_OD= 5.6583231 O_TOT= 110.7157112 <-- SPRD O_D= 6.0985789 O_OD= 4.4748109 O_TOT= 58.7709950 <-- SPRD O_D= 5.2660800 O_OD= 4.5811695 O_TOT= 58.0448547 <-- SPRD O_D= 5.0563240 O_OD= 4.5572975 O_TOT= 57.8112268 <-- SPRD O_D= 4.9553356 O_OD= 4.5798472 O_TOT= 57.7327880 <-- SPRD O_D= 4.8658124 O_OD= 4.5148747 O_TOT= 57.5782923 <-- SPRD O_D= 4.7876181 O_OD= 4.4716926 O_TOT= 57.4569159 <-- SPRD O_D= 4.7836031 O_OD= 4.5399151 O_TOT= 57.5211234 <-- SPRD O_D= 4.4656623 O_OD= 4.4629704 O_TOT= 57.1262379 <-- SPRD O_D= 4.1793155 O_OD= 4.6150183 O_TOT= 56.9919390 <-- SPRD O_D= 3.6857451 O_OD= 4.6589799 O_TOT= 56.5423303 <-- SPRD user at shell:~/Material/cubic/final_cal2/subdir_2$ grep CONV subdir_2.wout +--------------------------------------------------------------------+<-- CONV | Iter Delta Spread RMS Gradient Spread (Ang^2) Time |<-- CONV +--------------------------------------------------------------------+<-- CONV 0 0.931E+02 0.0000000000 93.0933355527 0.43 <-- CONV 1 0.176E+02 *************** 110.7157112167 0.46 <-- CONV 100 -0.153E-02 2.7516286420 58.7709949994 2.91 <-- CONV 200 -0.212E-02 2.3972358432 58.0448546806 5.39 <-- CONV 300 -0.342E-02 6.1861505188 57.8112267821 7.87 <-- CONV 400 -0.292E-02 1.7626116348 57.7327880169 10.36 <-- CONV 500 -0.109E-02 4.0748162263 57.5782923243 12.84 <-- CONV 600 -0.125E-02 2.0883925071 57.4569159120 15.32 <-- CONV 700 -0.567E-01 7.0512707925 57.5211233808 17.81 <-- CONV 800 -0.427E-03 1.4032648663 57.1262378986 20.29 <-- CONV 900 -0.773E-02 3.6088829215 56.9919390298 22.15 <-- CONV 1000 -0.115E-02 4.6967282749 56.5423302753 23.68 <-- CONV -------------- next part -------------- An HTML attachment was scrubbed... URL: From dwz4587 at 163.com Fri Nov 8 10:34:37 2013 From: dwz4587 at 163.com (dwz) Date: Fri, 8 Nov 2013 17:34:37 +0800 (CST) Subject: [Wannier] Ask for help about wannier90 Message-ID: <7f100db4.18c12.14237103226.Coremail.dwz4587@163.com> To those who may care : Hi there, the wannier-request system told me that I can mail you to post , I am not sure whether it is a proper way to ask for help of debug . If not ,it will be appreciated ur telling me the right way. I am using the Vasp2wannier function , which is based on the link between w90 and recompiled vasp, LWANNIER90_RUN tag in Vasp's INCAR file boot wannier_run and generate the wannier.min file, which is consistent with the CONTCAR and KPOINTS file . But it is known that VASP is unable to write PROJETION blocks in the wannier.win file , so I have to add it mannually . But rerun of VASP(set LWANNIER90 = .TRUE in INCAR this time ) with projection included wannier.win always fails . The wannier.wout reports : param_get_projections: Fewer projections defined than the number of Wannier functions requested . Projection block is necessary for the generation fo AMM file . So ,how to handle this problem ? Any response will be appreciated . Best wishes Yours -------------- next part -------------- An HTML attachment was scrubbed... URL: From pikuwww.mondal at gmail.com Fri Nov 8 14:16:37 2013 From: pikuwww.mondal at gmail.com (Wasim Mondal) Date: Fri, 8 Nov 2013 18:46:37 +0530 Subject: [Wannier] convergence problem Message-ID: Dear Wannier90 experts I am doing LDA+DMFT calculation of cubic peroskite system. I am using wien2wannier for generating Hamiltonian for DMFT calculation. After running the wien2wannier interface I have genarated Hamiltonian but I am not getting converged result. system is BaRuO3 in cubic phase. My Ru t2g and eg bands are in the energy range 5 to -2 eV which I have taken for the projection but not getting the converged result. please give some suggestion how I can get converged result. I am givng win file and convergence of the output. Please help me... iprint = 3 num_bands = 5 num_wann = 5 num_iter = 10000 num_print_cycles =500 !conv_window = 20 !conv_tol = 0.0001 !conv_noise_amp = 1 !conv_noise_num = 3 !dis_froz_min = 7. !dis_froz_max = 9. dis_mix_ratio = 0.5 write_proj = .true. write_xyz = .true. translate_home_cell = .true. !SYSTEM begin unit_cell_cart ang 7.5700950 0.0000000 0.0000000 0.0000000 7.5700950 0.0000000 0.0000000 0.0000000 7.5700950 end unit_cell_cart begin atoms_cart Ba 2.00296 2.00296 2.00296 Ru 0.00000 0.00000 0.00000 O 0.00000 0.00000 2.00296 O 2.00296 0.00000 0.00000 O 0.00000 2.00296 0.00000 end atoms_cart begin projections end projections begin kpoint_path end kpoint_path bands_plot = .false. wannier_plot = .false. !restart = plot hr_plot = .true. !dist_cutoff = 10 kmesh_tol = 0.0001 ! KPOINTS mp_grid : 8 8 8 begin kpoints 0.000000000 0.000000000 0.000000000 0.000000000 0.000000000 0.125000000 0.000000000 0.000000000 0.250000000 0.000000000 0.000000000 0.375000000 0.000000000 0.000000000 0.500000000 0.000000000 0.000000000 0.625000000 0.000000000 0.000000000 0.750000000 0.000000000 0.000000000 0.875000000 0.000000000 0.125000000 0.000000000 0.000000000 0.125000000 0.125000000 0.000000000 0.125000000 0.250000000 0.000000000 0.125000000 0.375000000 0.000000000 0.125000000 0.500000000 0.000000000 0.125000000 0.625000000 0.000000000 0.125000000 0.750000000 0.000000000 0.125000000 0.875000000 0.000000000 0.250000000 0.000000000 0.000000000 0.250000000 0.125000000 0.000000000 0.250000000 0.250000000 0.000000000 0.250000000 0.375000000 0.000000000 0.250000000 0.500000000 0.000000000 0.250000000 0.625000000 0.000000000 0.250000000 0.750000000 0.000000000 0.250000000 0.875000000 0.000000000 0.375000000 0.000000000 0.000000000 0.375000000 0.125000000 0.000000000 0.375000000 0.250000000 0.000000000 0.375000000 0.375000000 0.000000000 0.375000000 0.500000000 0.000000000 0.375000000 0.625000000 0.000000000 0.375000000 0.750000000 0.750000000 0.750000000 0.250000000 0.750000000 0.750000000 0.375000000 0.750000000 0.750000000 0.500000000 0.750000000 0.750000000 0.625000000 0.750000000 0.750000000 0.750000000 0.750000000 0.750000000 0.875000000 0.750000000 0.875000000 0.000000000 0.750000000 0.875000000 0.125000000 0.750000000 0.875000000 0.250000000 0.750000000 0.875000000 0.375000000 0.750000000 0.875000000 0.500000000 0.750000000 0.875000000 0.625000000 0.750000000 0.875000000 0.750000000 0.750000000 0.875000000 0.875000000 0.875000000 0.000000000 0.000000000 0.875000000 0.000000000 0.125000000 0.875000000 0.000000000 0.250000000 0.875000000 0.000000000 0.375000000 0.875000000 0.000000000 0.500000000 0.875000000 0.000000000 0.625000000 0.875000000 0.000000000 0.750000000 0.875000000 0.000000000 0.875000000 0.875000000 0.125000000 0.000000000 0.875000000 0.125000000 0.125000000 0.875000000 0.125000000 0.250000000 0.875000000 0.125000000 0.375000000 0.875000000 0.125000000 0.500000000 0.875000000 0.125000000 0.625000000 0.875000000 0.125000000 0.750000000 0.875000000 0.125000000 0.875000000 0.875000000 0.250000000 0.000000000 0.875000000 0.250000000 0.125000000 0.875000000 0.250000000 0.250000000 0.875000000 0.250000000 0.375000000 0.875000000 0.250000000 0.500000000 0.875000000 0.250000000 0.625000000 0.875000000 0.250000000 0.750000000 0.875000000 0.250000000 0.875000000 0.875000000 0.375000000 0.000000000 0.875000000 0.375000000 0.125000000 0.875000000 0.375000000 0.250000000 0.875000000 0.375000000 0.375000000 0.875000000 0.375000000 0.500000000 0.875000000 0.375000000 0.625000000 0.875000000 0.375000000 0.750000000 0.875000000 0.375000000 0.875000000 0.875000000 0.500000000 0.000000000 0.875000000 0.500000000 0.125000000 0.875000000 0.500000000 0.250000000 0.875000000 0.500000000 0.375000000 0.875000000 0.500000000 0.500000000 0.875000000 0.500000000 0.625000000 0.000000000 0.375000000 0.875000000 0.000000000 0.500000000 0.000000000 0.000000000 0.500000000 0.125000000 0.000000000 0.500000000 0.250000000 0.000000000 0.500000000 0.375000000 0.000000000 0.500000000 0.500000000 0.000000000 0.500000000 0.625000000 0.000000000 0.500000000 0.750000000 0.000000000 0.500000000 0.875000000 0.000000000 0.625000000 0.000000000 0.000000000 0.625000000 0.125000000 0.000000000 0.625000000 0.250000000 0.000000000 0.625000000 0.375000000 0.000000000 0.625000000 0.500000000 0.000000000 0.625000000 0.625000000 0.000000000 0.625000000 0.750000000 end kpoint_path grep CONV subdir_final3.wout +--------------------------------------------------------------------+<-- CONV | Iter Delta Spread RMS Gradient Spread (Ang^2) Time |<-- CONV +--------------------------------------------------------------------+<-- CONV 0 0.137E+03 0.0000000000 137.4045564549 0.60 <-- CONV 1 0.160E+02 *************** 153.3761552128 0.65 <-- CONV 500 -0.779E-04 1.3607653529 88.5601454505 16.17 <-- CONV 1000 -0.109E-01 4.1991000285 88.1893434459 31.55 <-- CONV 1500 -0.164E+00 9.9892858216 89.0180260346 46.93 <-- CONV 2000 -0.207E-02 5.0277047303 88.0187200928 62.30 <-- CONV 2500 -0.149E-01 8.1926771352 87.3785199703 77.68 <-- CONV 3000 0.691E+00 207.5630284831 88.2665159542 93.07 <-- CONV 3500 -0.173E-03 1.9355601893 87.0893426503 108.46 <-- CONV 4000 -0.250E-01 4.3581029673 86.7670537954 123.87 <-- CONV 4500 -0.301E-03 2.1800252240 86.3232664973 139.30 <-- CONV 5000 0.171E+00 21.7928277599 86.5733258404 154.74 <-- CONV 5500 -0.169E-01 3.2675143541 86.3221489576 170.18 <-- CONV 6000 -0.131E-01 4.3164974607 87.8598594553 185.62 <-- CONV 6500 -0.218E-02 4.4854761339 86.6882760134 201.06 <-- CONV 7000 -0.177E-02 5.4546515036 86.3730631479 216.52 <-- CONV 7500 -0.219E-02 4.5207429232 86.2131022775 231.98 <-- CONV 8000 -0.435E-02 5.3779798218 86.1651066041 247.44 <-- CONV 8500 -0.698E-04 7.2862485116 86.2021739674 262.91 <-- CONV 9000 -0.132E-01 4.9740400263 86.3604114617 278.37 <-- CONV 9500 -0.483E-01 9.7638010532 86.2606441552 293.82 <-- CONV 10000 -0.703E-02 14.0944955193 86.6690086165 309.27 <-- CONV -------------- next part -------------- An HTML attachment was scrubbed... URL: From ehs73 at cornell.edu Fri Nov 8 17:41:52 2013 From: ehs73 at cornell.edu (E. H. Smith) Date: Fri, 8 Nov 2013 11:41:52 -0500 Subject: [Wannier] Ask for help about wannier90 In-Reply-To: <7f100db4.18c12.14237103226.Coremail.dwz4587@163.com> References: <7f100db4.18c12.14237103226.Coremail.dwz4587@163.com> Message-ID: Follow this procedure: 1. Make a directory with your VASP input files. Run VASP with LWANNIER90=.TRUE. 2. Modify with seedname.win file (which will have been written for you). For some reason I don't understand, the seedname.win file is written slightly incorrectly. You have to change num_wann = 32 !where I use 32 as an example; this value will be written by VASP to num_bands= 32 !i.e. the number that was written for num_wann num_wann = N !for N the number of projections you want. e.g. if you want the p projectors for a two-atom basis, num_wann would be 6. Continue modifying seedname.win with any other parameters you want. 3. Continue with wannier90 calculations. Cheers. On Fri, Nov 8, 2013 at 4:34 AM, dwz wrote: > To those who may care : > Hi there, the wannier-request system told me that I can mail you to post , > I am not sure whether it is a proper way to ask for help of debug . If not > ,it will be appreciated ur telling me the right way. > I am using the Vasp2wannier function , which is based on the link between > w90 and recompiled vasp, *LWANNIER90_RUN* tag in Vasp's INCAR file boot > wannier_run and generate the wannier.min file, which is consistent with > the CONTCAR and KPOINTS file . But it is known that VASP is unable to write > PROJETION blocks in the wannier.win file , so I have to add it mannually . > But rerun of VASP(set *LWANNIER90* = .TRUE in INCAR this time ) with > projection included wannier.win always fails . The wannier.wout reports : *param_get_projections: > Fewer projections defined than the number of Wannier functions requested . * > Projection block is necessary for the generation fo AMM file . > So ,how to handle this problem ? > Any response will be appreciated . > > Best wishes > Yours > > > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From brehmj at sas.upenn.edu Fri Nov 8 23:21:56 2013 From: brehmj at sas.upenn.edu (JOHN BREHM) Date: Fri, 8 Nov 2013 17:21:56 -0500 Subject: [Wannier] failire constantly doing pw2wannier Message-ID: using the following 2 lines from my run script, the pw2wannier.x runs but fails: export ESPRESSO_NPROCS=64 pw2wannier90.x pw2wan.out here is the input file l0.pw2wan: &inputpp outdir = './' prefix = 'pwscf' seedname = 'l0' spin_component = 'none' write_mmn = .true. write_amn = .true. write_unk = .true. / I ran scf, nscf, and wannier.x -pp all with the same number of kpoints. my wannier -pp generates the file l0.nnkp. here are the last few lines of pw2wan.out: Spin CASE ( default = unpolarized ) Wannier mode is: standalone ----------------- *** Reading nnkp ----------------- Checking info from wannier.nnkp file Something wrong! rlatt(i,j) = 1.88972613288564 at(i,j)= 1.00000000000000 would someone be able to tell me what i am doing wrong? tnx, jb -------------- next part -------------- An HTML attachment was scrubbed... URL: From giovanni.pizzi at epfl.ch Fri Nov 8 23:46:06 2013 From: giovanni.pizzi at epfl.ch (Giovanni Pizzi) Date: Fri, 8 Nov 2013 22:46:06 +0000 Subject: [Wannier] failire constantly doing pw2wannier In-Reply-To: References: Message-ID: <9E730F6D-DD44-4877-B420-74E3AE5DCBA3@epfl.ch> Dear John, first of all remember to sign your posts with your affiliation. Regarding your problem, the number 1.88972613288564 in the error message looks suspiciously like the bohr<->angstrom conversion. Remember that default units (if nothing else is specified) are bohr in QE, and angstrom in Wannier90 (but it is possible to switch to bohr, see for instance section 2.4.3 of the manual of Wannier90 2.0 here: http://www.wannier.org/doc/user_guide.pdf - or alternatively use angstrom in QE). Anyway, it is hard to tell more without the input files you used for QE and Wannier90. Please double check the definition of the unit cell and of the atomic positions in all your input files. Best, Giovanni Pizzi -- Giovanni Pizzi Post-doctoral Research Scientist EPFL STI IMX THEOS MXC 340 (B?timent MXC) Station 12 CH-1015 Lausanne (Switzerland) Phone: +41 21 69 31124 On 8 Nov 2013, at 23:21, JOHN BREHM wrote: using the following 2 lines from my run script, the pw2wannier.x runs but fails: export ESPRESSO_NPROCS=64 pw2wannier90.x pw2wan.out here is the input file l0.pw2wan: &inputpp outdir = './' prefix = 'pwscf' seedname = 'l0' spin_component = 'none' write_mmn = .true. write_amn = .true. write_unk = .true. / I ran scf, nscf, and wannier.x -pp all with the same number of kpoints. my wannier -pp generates the file l0.nnkp. here are the last few lines of pw2wan.out: Spin CASE ( default = unpolarized ) Wannier mode is: standalone ----------------- *** Reading nnkp ----------------- Checking info from wannier.nnkp file Something wrong! rlatt(i,j) = 1.88972613288564 at(i,j)= 1.00000000000000 would someone be able to tell me what i am doing wrong? tnx, jb _______________________________________________ Wannier mailing list Wannier at quantum-espresso.org http://www.democritos.it/mailman/listinfo/wannier -------------- next part -------------- An HTML attachment was scrubbed... URL: From a.mostofi at imperial.ac.uk Sat Nov 9 10:26:13 2013 From: a.mostofi at imperial.ac.uk (Arash Mostofi) Date: Sat, 09 Nov 2013 09:26:13 +0000 Subject: [Wannier] convergence problem In-Reply-To: References: Message-ID: <527DFFB5.2090700@imperial.ac.uk> Dear Wasim, Thank you for your post to the mailing list. Before I get to your question, may I just remind you that we ask everyone on the mailing list to provide their name *and* affiliation in their posts - it is a simple thing to do. Regarding your calculation for BaRuO3, have you checked the bandstructure to see whether the t2g/eg bands that you are trying to Wannierize are completely isolated from all other states in the Brillouin zone? If not, then you will need to disentangle them. You can find more information about disentanglement in Section 2.I of Rev Mod Phys 84, 1419 (2012) and, in more detail, in the original article on dealing with entangled bands, Phys Rev B 65 035109 (2001). In any case, if you know what WFs you are looking for, it is a good idea to define specific projections in the "projections" block of the input file: leaving that block empty as you have done results in the use of randomly placed spherical orbitals as the initial guess for the WFs, which is not an optimal starting point for your system. Best wishes, Arash -- Dr Arash A Mostofi Departments of Materials and Physics Deputy Director, CDT on Theory and Simulation of Materials Imperial College London, London SW7 2AZ, UK +44 (0)207 594 8154 | www.cmth.ph.ic.ac.uk/people/a.mostofi On 08/11/2013 13:16, Wasim Mondal wrote: > Dear Wannier90 experts > I am doing LDA+DMFT calculation of cubic > peroskite system. I am using wien2wannier for generating Hamiltonian for > DMFT calculation. After running the wien2wannier interface I have > genarated Hamiltonian but I am not getting converged result. system is > BaRuO3 in cubic phase. My Ru t2g and eg bands are in the energy range 5 > to -2 eV which I have taken for the projection but not getting the > converged result. please give some suggestion how I can get converged > result. I am givng win file and convergence of the output. Please help me... > > iprint = 3 > num_bands = 5 > num_wann = 5 > num_iter = 10000 > num_print_cycles =500 > !conv_window = 20 > !conv_tol = 0.0001 > !conv_noise_amp = 1 > !conv_noise_num = 3 > > !dis_froz_min = 7. > !dis_froz_max = 9. > dis_mix_ratio = 0.5 > write_proj = .true. > write_xyz = .true. > translate_home_cell = .true. > > !SYSTEM > > begin unit_cell_cart > ang > 7.5700950 0.0000000 0.0000000 > 0.0000000 7.5700950 0.0000000 > 0.0000000 0.0000000 7.5700950 > end unit_cell_cart > > begin atoms_cart > Ba 2.00296 2.00296 2.00296 > Ru 0.00000 0.00000 0.00000 > O 0.00000 0.00000 2.00296 > O 2.00296 0.00000 0.00000 > O 0.00000 2.00296 0.00000 > end atoms_cart > > begin projections > end projections > > begin kpoint_path > end kpoint_path > > bands_plot = .false. > wannier_plot = .false. > !restart = plot > hr_plot = .true. > !dist_cutoff = 10 > > kmesh_tol = 0.0001 > > ! KPOINTS > > mp_grid : 8 8 8 > > begin kpoints > 0.000000000 0.000000000 0.000000000 > 0.000000000 0.000000000 0.125000000 > 0.000000000 0.000000000 0.250000000 > 0.000000000 0.000000000 0.375000000 > 0.000000000 0.000000000 0.500000000 > 0.000000000 0.000000000 0.625000000 > 0.000000000 0.000000000 0.750000000 > 0.000000000 0.000000000 0.875000000 > 0.000000000 0.125000000 0.000000000 > 0.000000000 0.125000000 0.125000000 > 0.000000000 0.125000000 0.250000000 > 0.000000000 0.125000000 0.375000000 > 0.000000000 0.125000000 0.500000000 > 0.000000000 0.125000000 0.625000000 > 0.000000000 0.125000000 0.750000000 > 0.000000000 0.125000000 0.875000000 > 0.000000000 0.250000000 0.000000000 > 0.000000000 0.250000000 0.125000000 > 0.000000000 0.250000000 0.250000000 > 0.000000000 0.250000000 0.375000000 > 0.000000000 0.250000000 0.500000000 > 0.000000000 0.250000000 0.625000000 > 0.000000000 0.250000000 0.750000000 > 0.000000000 0.250000000 0.875000000 > 0.000000000 0.375000000 0.000000000 > 0.000000000 0.375000000 0.125000000 > 0.000000000 0.375000000 0.250000000 > 0.000000000 0.375000000 0.375000000 > 0.000000000 0.375000000 0.500000000 > 0.000000000 0.375000000 0.625000000 > 0.000000000 0.375000000 0.750000000 > 0.750000000 0.750000000 0.250000000 > 0.750000000 0.750000000 0.375000000 > 0.750000000 0.750000000 0.500000000 > 0.750000000 0.750000000 0.625000000 > 0.750000000 0.750000000 0.750000000 > 0.750000000 0.750000000 0.875000000 > 0.750000000 0.875000000 0.000000000 > 0.750000000 0.875000000 0.125000000 > 0.750000000 0.875000000 0.250000000 > 0.750000000 0.875000000 0.375000000 > 0.750000000 0.875000000 0.500000000 > 0.750000000 0.875000000 0.625000000 > 0.750000000 0.875000000 0.750000000 > 0.750000000 0.875000000 0.875000000 > 0.875000000 0.000000000 0.000000000 > 0.875000000 0.000000000 0.125000000 > 0.875000000 0.000000000 0.250000000 > 0.875000000 0.000000000 0.375000000 > 0.875000000 0.000000000 0.500000000 > 0.875000000 0.000000000 0.625000000 > 0.875000000 0.000000000 0.750000000 > 0.875000000 0.000000000 0.875000000 > 0.875000000 0.125000000 0.000000000 > 0.875000000 0.125000000 0.125000000 > 0.875000000 0.125000000 0.250000000 > 0.875000000 0.125000000 0.375000000 > 0.875000000 0.125000000 0.500000000 > 0.875000000 0.125000000 0.625000000 > 0.875000000 0.125000000 0.750000000 > 0.875000000 0.125000000 0.875000000 > 0.875000000 0.250000000 0.000000000 > 0.875000000 0.250000000 0.125000000 > 0.875000000 0.250000000 0.250000000 > 0.875000000 0.250000000 0.375000000 > 0.875000000 0.250000000 0.500000000 > 0.875000000 0.250000000 0.625000000 > 0.875000000 0.250000000 0.750000000 > 0.875000000 0.250000000 0.875000000 > 0.875000000 0.375000000 0.000000000 > 0.875000000 0.375000000 0.125000000 > 0.875000000 0.375000000 0.250000000 > 0.875000000 0.375000000 0.375000000 > 0.875000000 0.375000000 0.500000000 > 0.875000000 0.375000000 0.625000000 > 0.875000000 0.375000000 0.750000000 > 0.875000000 0.375000000 0.875000000 > 0.875000000 0.500000000 0.000000000 > 0.875000000 0.500000000 0.125000000 > 0.875000000 0.500000000 0.250000000 > 0.875000000 0.500000000 0.375000000 > 0.875000000 0.500000000 0.500000000 > 0.875000000 0.500000000 0.625000000 > 0.000000000 0.375000000 0.875000000 > 0.000000000 0.500000000 0.000000000 > 0.000000000 0.500000000 0.125000000 > 0.000000000 0.500000000 0.250000000 > 0.000000000 0.500000000 0.375000000 > 0.000000000 0.500000000 0.500000000 > 0.000000000 0.500000000 0.625000000 > 0.000000000 0.500000000 0.750000000 > 0.000000000 0.500000000 0.875000000 > 0.000000000 0.625000000 0.000000000 > 0.000000000 0.625000000 0.125000000 > 0.000000000 0.625000000 0.250000000 > 0.000000000 0.625000000 0.375000000 > 0.000000000 0.625000000 0.500000000 > 0.000000000 0.625000000 0.625000000 > 0.000000000 0.625000000 0.750000000 > > end kpoint_path > > grep CONV subdir_final3.wout > +--------------------------------------------------------------------+<-- CONV > | Iter Delta Spread RMS Gradient Spread (Ang^2) Time > |<-- CONV > +--------------------------------------------------------------------+<-- CONV > 0 0.137E+03 0.0000000000 137.4045564549 0.60 > <-- CONV > 1 0.160E+02 *************** 153.3761552128 0.65 > <-- CONV > 500 -0.779E-04 1.3607653529 88.5601454505 16.17 > <-- CONV > 1000 -0.109E-01 4.1991000285 88.1893434459 31.55 > <-- CONV > 1500 -0.164E+00 9.9892858216 89.0180260346 46.93 > <-- CONV > 2000 -0.207E-02 5.0277047303 88.0187200928 62.30 > <-- CONV > 2500 -0.149E-01 8.1926771352 87.3785199703 77.68 > <-- CONV > 3000 0.691E+00 207.5630284831 88.2665159542 93.07 > <-- CONV > 3500 -0.173E-03 1.9355601893 87.0893426503 108.46 > <-- CONV > 4000 -0.250E-01 4.3581029673 86.7670537954 123.87 > <-- CONV > 4500 -0.301E-03 2.1800252240 86.3232664973 139.30 > <-- CONV > 5000 0.171E+00 21.7928277599 86.5733258404 154.74 > <-- CONV > 5500 -0.169E-01 3.2675143541 86.3221489576 170.18 > <-- CONV > 6000 -0.131E-01 4.3164974607 87.8598594553 185.62 > <-- CONV > 6500 -0.218E-02 4.4854761339 86.6882760134 201.06 > <-- CONV > 7000 -0.177E-02 5.4546515036 86.3730631479 216.52 > <-- CONV > 7500 -0.219E-02 4.5207429232 86.2131022775 231.98 > <-- CONV > 8000 -0.435E-02 5.3779798218 86.1651066041 247.44 > <-- CONV > 8500 -0.698E-04 7.2862485116 86.2021739674 262.91 > <-- CONV > 9000 -0.132E-01 4.9740400263 86.3604114617 278.37 > <-- CONV > 9500 -0.483E-01 9.7638010532 86.2606441552 293.82 > <-- CONV > 10000 -0.703E-02 14.0944955193 86.6690086165 309.27 > <-- CONV > > > > > > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > From brehmj at sas.upenn.edu Sat Nov 9 20:19:41 2013 From: brehmj at sas.upenn.edu (JOHN BREHM) Date: Sat, 9 Nov 2013 14:19:41 -0500 Subject: [Wannier] error message in pw2wannier run Message-ID: hi everyone. has anyone else ever seen this error when running pw2wannier: *** Reading nnkp ----------------- Checking info from wannier.nnkp file - Real lattice is ok - Reciprocal lattice is ok Something wrong! numk= 256 iknum= 424 I checked bothj QE and wannier90 tutorial and useer guide, but can't find either numk or iknum. that said, i am using 256 kpoints. tnx, John Brehm Univ.of Pennsylvania -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Sat Nov 9 20:46:47 2013 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Sat, 9 Nov 2013 19:46:47 +0000 Subject: [Wannier] error message in pw2wannier run In-Reply-To: References: Message-ID: <78877D222077A642A4CFAB09BD0077E026FE1866@MBX10.ad.oak.ox.ac.uk> On 9 Nov 2013, at 19:19, JOHN BREHM wrote: > hi everyone. > > has anyone else ever seen this error when running pw2wannier: > > *** Reading nnkp > ----------------- > > Checking info from wannier.nnkp file > > - Real lattice is ok > - Reciprocal lattice is ok > Something wrong! > numk= 256 iknum= 424 > > > I checked bothj QE and wannier90 tutorial and useer guide, but can't find either numk or iknum. These are internal variables in pw2wannier90.f90. Admittedly the error messages in pw2wannier90 are not as friendly as the ones in wannnier90 itself (at least I hope they are friendly and informative - please tell us if you find one that isn't). numk is the number of kpoints found in the nnkp file. So this is as you expected iknum is the number of kpoints in the pwscf run - which is > 256. If you look at the output of your nscf run you should see it has 424 kpoints. I suspect this is due to using symmetry - try setting nosym=.true. in the system namelist and rerunning the nscf calculation. Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From amirhosseinba67 at gmail.com Tue Nov 12 22:13:51 2013 From: amirhosseinba67 at gmail.com (Amir hossein) Date: Wed, 13 Nov 2013 00:43:51 +0330 Subject: [Wannier] from kpoints_init : error # 21888313 Message-ID: Hi Dear all I did scf,nscf,dft_bands,pwexport,disentangle and wannier for Si in test01 from tests of want-2.4.0, when I want to perform bands.in, the program interrupt and in bands.out file this error is appear: "from kpoints_init : error # 21888313 kpt grid not Monkhorst-Pack" What am I doing? Thanks to all From dwz4587 at 163.com Thu Nov 14 03:51:22 2013 From: dwz4587 at 163.com (dwz) Date: Thu, 14 Nov 2013 10:51:22 +0800 (CST) Subject: [Wannier] Error reported when running postw90.x Message-ID: <6e8daf37.11bd9.14254852941.Coremail.dwz4587@163.com> Dear Wannier developers & experts I am using vasp2wannier , but when running postw90.x , Error: Problem opening input file wannier90.spn . Here is my work : 1. Run vasp with LWANNIER90=.TRUE. -tag in INCAR and the vasp2wannier function will yield the wannier.win wannier.amn wannier.mmn and wannier.wout files. 2. Run wannier to cal the MLWF 3. Add some Berry curvature plots parameters in wannier.win as example18 in the Wannier tutorial book , reading fermi_energy = [insert your value here] berry_curv_unit = bohr2 kpath = true kpath_task = bands+curv kpath_bands_colour = spin kpath_num_points = 1000 4. Run postw90.x . Then is my bug report in wannier.wpout file Error: Problem opening input file wannier90.spn . Since most calculations in official wannier mtrls are performed by pw2wannier , the lack of vasp2wannier info make my situation really murky .So, how to avoid postw.x reading wannier90.spn ? If the spn file is really necessary , how to generate it ? It seems that pw2wannier can generate this file ,does vasp2wannier function also do (if so , should I add some extra tag in INCAR file ) ? Any response will be appreciated Best ! -------------- next part -------------- An HTML attachment was scrubbed... URL: From giovanni.pizzi at epfl.ch Thu Nov 14 14:12:58 2013 From: giovanni.pizzi at epfl.ch (Giovanni Pizzi) Date: Thu, 14 Nov 2013 14:12:58 +0100 Subject: [Wannier] Error reported when running postw90.x In-Reply-To: <6e8daf37.11bd9.14254852941.Coremail.dwz4587@163.com> References: <6e8daf37.11bd9.14254852941.Coremail.dwz4587@163.com> Message-ID: <5284CC5A.7000708@epfl.ch> Dear dwz, first of all I remind you that you should sign your emails with your name and affiliation. Regarding your problem: some modules of postw90.x need to have information about the spin (e.g. the berry and orbital magnetization parts, examples 18-19). Other modules of postw90.x instead can work also without this information (see e.g. example 17). Therefore, if you are interested in one of the applications that require this additional spin information, the .spn file is required. Since the additional features were added only one month ago, when Wannier90 version 2.0 came out, it is very probable that most interfaces are still not able to produce this file. Since one of the codes that the Wannier90 developers use is Quantum ESPRESSO, we provided the implementation for QE in the pw2wannier90.x interface, but it is indeed very probable that the VASP interface does not have this feature yet. What you can do is to try to ask the VASP2WANNIER developers if they are willing to introduce the new features in the code, if they did not do it yet. Best, Giovanni Pizzi On 11/14/2013 03:51 AM, dwz wrote: > Dear Wannier developers & experts > I am using vasp2wannier , but when running postw90.x , Error: Problem > opening input file wannier90.spn . > Here is my work : > 1. Run vasp with LWANNIER90=.TRUE. -tag in INCAR and the vasp2wannier > function will yield the wannier.win wannier.amn wannier.mmn and > wannier.wout files. > 2. Run wannier to cal the MLWF > 3. Add some Berry curvature plots parameters in wannier.win as > example18 in the Wannier tutorial book , reading > fermi_energy = [insert your value here] > berry_curv_unit = bohr2 > kpath = true > kpath_task = bands+curv > kpath_bands_colour = spin > kpath_num_points = 1000 > 4. Run postw90.x . Then is my bug report in wannier.wpout file Error: > Problem opening input file wannier90.spn . > Since most calculations in official wannier mtrls are performed by > pw2wannier , the lack of vasp2wannier info make my situation really > murky .So, how to avoid postw.x reading wannier90.spn ? If the spn > file is really necessary , how to generate it ? It seems that > pw2wannier can generate this file ,does vasp2wannier function also do > (if so , should I add some extra tag in INCAR file ) ? > Any response will be appreciated > Best ! > > > > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier -- Giovanni Pizzi Post-doctoral Research Scientist EPFL STI IMX THEOS MXC 340 (B?timent MXC) Station 12 CH-1015 Lausanne (Switzerland) Phone: +41 21 69 31124 -------------- next part -------------- An HTML attachment was scrubbed... URL: From rio10002 at gmail.com Thu Nov 14 18:30:07 2013 From: rio10002 at gmail.com (satrio) Date: Thu, 14 Nov 2013 12:30:07 -0500 Subject: [Wannier] SOC calculation using Wannier90 Message-ID: Dear All I tried to do calculation for MoS2 using spin orbit coupling (SOC). So first i do DFT calculation using pw.x Quantum espresso. Then i run pw2wannier90.x. But it gave me error when i run pw2wannier90.x. It seems it had a problem in pseudopotential that i used. In first trial, i use Ultrasoft and full relativistic pseudopotential. When i run pw2wannier90.x, i got the error like this from pw2wannier90 : error # 1 NCLS calculation not implimented with USP Then i change my pseudopotential Second trial I used PAW and full relativistic pseudopotential. But it still gave me error when i run pw2wannier90.x. The error said from PAW_gcxc_v : error # 1 Non collinear not implemented So i really dont know how to fix this. Those are two kind of pseudopetential that are available in quantum espresso that are full relativistic. Is there any suggestion for me? Thank you very much Regards Rio -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Thu Nov 14 18:35:05 2013 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Thu, 14 Nov 2013 17:35:05 +0000 Subject: [Wannier] SOC calculation using Wannier90 In-Reply-To: References: Message-ID: <78877D222077A642A4CFAB09BD0077E027020A70@MBX10.ad.oak.ox.ac.uk> On 14 Nov 2013, at 17:30, satrio wrote: > Dear All > > I tried to do calculation for MoS2 using spin orbit coupling (SOC). So first i do DFT calculation using pw.x Quantum espresso. Then i run pw2wannier90.x. But it gave me error when i run pw2wannier90.x. It seems it had a problem in pseudopotential that i used. > In first trial, i use Ultrasoft and full relativistic pseudopotential. When i run pw2wannier90.x, i got the error like this > from pw2wannier90 : error # 1 > NCLS calculation not implimented with USP > > Then i change my pseudopotential > > Second trial I used PAW and full relativistic pseudopotential. But it still gave me error when i run pw2wannier90.x. The error said > from PAW_gcxc_v : error # 1 > Non collinear not implemented > > So i really dont know how to fix this. Those are two kind of pseudopetential that are available in quantum espresso that are full relativistic. Is there any suggestion for me? Computing the correct overlaps for USP/PAW and spinor wave functions has not been implemented in pw2wannier90. There is no fundamental reason, just that no one has coded it yet. You need to use norm-conserving potentials with spin-orbit coupling. QE's ld1 code is capable of generating such potentials. Yours Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From creaman1986 at gmail.com Thu Nov 14 18:50:43 2013 From: creaman1986 at gmail.com (=?Big5?B?rFK8/bOv?=) Date: Fri, 15 Nov 2013 01:50:43 +0800 Subject: [Wannier] about disentanglement Message-ID: Hello everyone I have a little question about disentanglement which confused me for a while. After the disentanglement procedure ,the program will pick up some appropriate states to construct Wannier function. So I suppose "Eigenvalues inside optimal subspace" should be equal to some of "Original eigenvalues inside outer window" , but I found they are not equal. So, is it ok to be "not equal"? if not, how can I reduce the error? thanks -------------- next part -------------- An HTML attachment was scrubbed... URL: From zgxolg at gmail.com Fri Nov 15 03:47:54 2013 From: zgxolg at gmail.com (Xiaolong Zhang) Date: Fri, 15 Nov 2013 10:47:54 +0800 Subject: [Wannier] Problem when generating Wannier functions using mixed non-relativistic and full relativistic PP Message-ID: Dear developers and users, I come across this problem when I try to study SiC system, where full relativistic PP of Si and non-relativistic PP of C are used. I can obtain perfect MLWFs in spinless case using 4 projections(sp3 orbitals of Si). But when the PP of SiI is substituted with a full relativistic one, the wannier90 code keeps complaining insufficient projections(the 4 sp3 projecitons now correspond an 8 projections in SOC calculation). It seems Q.E. (I use 4.0.3 ) treats this mixed PP job as an spinless one as the weight of irr.kpoint in scf calculation doesn't halve as I expected. If this is true, how can I proceed to generate correct MLWFs with mixed non-relativistic and full relativistic PP, any suggestions would be highly appreciated. Thank you in advance! Xiaolong -- Xiao-Long Zhang Ph.D student, Institute of Physics, Chinese Academy of Sciences 8 Nansanjie, Zhongguancun Beijing 100190, China Email: xlzhang at iphy.ac.cn OR zgxolg at gmail.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From elias.assmann at gmail.com Fri Nov 15 12:07:43 2013 From: elias.assmann at gmail.com (Elias Assmann) Date: Fri, 15 Nov 2013 12:07:43 +0100 Subject: [Wannier] guiding centers Message-ID: <5286007F.2050104@gmail.com> Hi List, I would like to understand how the ?guiding centers? that can be specified in the ?win? file work. I have checked the user guide, tutorial, and review article, but cannot find anything definite. Elias -- Elias Assmann Institute of Solid State Physics Vienna University of Technology From amirhosseinba67 at gmail.com Fri Nov 15 13:37:58 2013 From: amirhosseinba67 at gmail.com (Amir hossein) Date: Fri, 15 Nov 2013 16:07:58 +0330 Subject: [Wannier] kpt grid not Monkhorst-Pack Message-ID: Dear all I did scf,nscf,dft_bands,pwexport,disentangle and wannier for Si in test01 from tests of want-2.4.0, when I want to perform bands.in, the program interrupt and in bands.out file this error is appear: "from kpoints_init : error # 21888313 kpt grid not Monkhorst-Pack" What am I doing? Thanks to all From giovanni.pizzi at epfl.ch Fri Nov 15 13:42:03 2013 From: giovanni.pizzi at epfl.ch (Giovanni Pizzi) Date: Fri, 15 Nov 2013 13:42:03 +0100 Subject: [Wannier] kpt grid not Monkhorst-Pack In-Reply-To: References: Message-ID: <5286169B.3030501@epfl.ch> Dear Amir, this is the forum for the Wannier90 code, that is a different code with respect to WanT. For help regarding WanT, address to their mailing list. Best, Giovanni Pizzi On 11/15/2013 01:37 PM, Amir hossein wrote: > Dear all > > I did scf,nscf,dft_bands,pwexport,disentangle and wannier for Si in > test01 from tests of want-2.4.0, when I want to perform bands.in, the > program interrupt and in bands.out file this error is appear: > > "from kpoints_init : error # 21888313 > kpt grid not Monkhorst-Pack" > > What am I doing? > > > Thanks to all > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier -- Giovanni Pizzi Post-doctoral Research Scientist EPFL STI IMX THEOS MXC 340 (B?timent MXC) Station 12 CH-1015 Lausanne (Switzerland) Phone: +41 21 69 31124 From nicola.marzari at epfl.ch Fri Nov 15 14:57:07 2013 From: nicola.marzari at epfl.ch (Nicola Marzari) Date: Fri, 15 Nov 2013 14:57:07 +0100 Subject: [Wannier] guiding centers In-Reply-To: <5286007F.2050104@gmail.com> References: <5286007F.2050104@gmail.com> Message-ID: <52862833.2090503@epfl.ch> On 15/11/2013 12:07, Elias Assmann wrote: > Hi List, > > I would like to understand how the ?guiding centers? that can be > specified in the ?win? file work. I have checked the user guide, > tutorial, and review article, but cannot find anything definite. > > Elias > Dear Elias, the idea is that if M_nn_kb was dispersionless, it would be identical to b scalar rcenter, and (see eq 28 of the 2012 review) would give you rcenter. Since once the functions are maximally localized, M_nn is optimally smooth as you surf the Brillouin zone, you want to make sure that as you move around, the branch cuts are such that M_nn gives you a Im ln M_nn that is roughly always b scalar rguide, where rguide is an educated guess to rcenter. That info is still buried in the wann_phases subroutine inside wannierise.f90 - a bit more info in this old f77 version. nicola -- ---------------------------------------------------------------------- Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL -------------- next part -------------- c----------------------------------------------------------------------- subroutine phases(nbans,nkpts2,nntot,nnh,neigh, & bk,bka,cm,csheet,sheet,rguide,irguide) INCLUDE 'wannier.include' IMPLICIT COMPLEX (C) dimension cm(nbands,nbands,nkpts,nnmx), csum(nnmxh) dimension bk(3,nkpts,nnmx), bka(3,nnmxh), nntot(nkpts) dimension neigh(nkpts,nnmxh), rguide(3,nbands) dimension xx(nnmx), smat(3,3),svec(3),sinv(3,3) dimension csheet(nbands,nkpts,nnmx),sheet(nbands,nkpts,nnmx) dimension rnkb(nbands,nkpts,nnmx) ci=(0.,1.) c report problem to solve c for each band, csum is determined and then its appropriate c guiding center rguide(3,nwann) do nwann=1,nbans c get average phase for each unique bk direction do na=1,nnh csum(na)=(0.0,0.0) do nkp=1,nkpts2 nn=neigh(nkp,na) csum(na)=csum(na)+cm(nwann,nwann,nkp,nn) end do end do c now analyze that information to get good guess at c wannier center c write(*,*) c do na=1,nnh c write(*,'a,3f10.5,a,2f10.5)') c & ' bka=',(bka(j,na),j=1,3),' csum=',csum(na) c end do c problem is to find a real-space 3-vector rguide such that c phase of csum(nn) ~= phase of exp[ -i bka(nn) dot rguide ] c or, letting c xx(nn) = - Im ln csum(nn) (modulo 2*pi) c then c bka(nn) dot rguide ~= xx(nn) c c we take an arbitrary branch cut for first three xx(nn) c and determine rguide from these; then for each additional bka c vector, we first determine the most consistent branch cut, c and then update rguide c c rguide is obtained by minimizing c sum_nn [ bka(nn) dot rguide - xx(nn) ] ^2 c or, setting the derivative with respect to rcenter to zero, c sum_i smat(j,i) * rguide(i,nwann) = svec(j) c where c smat(j,i) = sum_nn bka(j,nn) * bka(i,nn) c svec(j) = sum_nn bka(j,nn) * xx(nn) c initialize smat and svec do j=1,3 do i=1,3 smat(j,i)=0. end do svec(j)=0. end do write(*,*) ' ' do nn=1,nnh if (nn.le.3) then c obtain xx with arbitrary branch cut choice xx(nn)=-dimag(log(csum(nn))) else c obtain xx with branch cut choice guided by rguide xx0=0. do j=1,3 xx0=xx0+bka(j,nn)*rguide(j,nwann) end do c xx0 is expected value for xx csumt=cexp(ci*xx0) c csumt has opposite of expected phase of csum(nn) xx(nn)=xx0-dimag(log(csum(nn)*csumt)) endif write(*,'(a,i5,3f7.3,2f10.5)') 'nn, bka, xx, mag =', 1 nn,(bka(j,nn),j=1,3),xx(nn),abs(csum(nn))/float(nkpts2) c update smat and svec do j=1,3 do i=1,3 smat(j,i)=smat(j,i)+bka(j,nn)*bka(i,nn) end do svec(j)=svec(j)+bka(j,nn)*xx(nn) end do if (nn.ge.3) then c determine rguide call inv3(smat,sinv,det) c the inverse of smat is sinv/det if (abs(det).lt.1.e-06) then c evidently first three bka vectors are linearly dependent c this is not allowed write(*,*) ' error in findr: dependency' stop endif if (irguide.ne.0) then do j=1,3 rguide(j,nwann)=0. do i=1,3 rguide(j,nwann)=rguide(j,nwann)+sinv(j,i)*svec(i)/det end do end do end if write(*,'(a,3f10.5)') 'rguide =',(rguide(i,nwann),i=1,3) endif end do end do c obtain branch cut choice guided by rguid do nkp=1,nkpts2 do nwann=1,nbans do nn=1,nntot(nkp) sheet(nwann,nkp,nn)=0.0 do j=1,3 sheet(nwann,nkp,nn)=sheet(nwann,nkp,nn)+ & bk(j,nkp,nn)*rguide(j,nwann) end do csheet(nwann,nkp,nn)=cexp(ci*sheet(nwann,nkp,nn)) end do end do end do c now check that we picked the proper sheet for the log c of cm. criterion: q_n^{k,b}=Im(ln(M_nn^{k,b})) + b \cdot r_n are c circa 0 for a good solution, circa multiples of 2 pi for a bad one. c I use the guiding center, instead of r_n, to understand which could be c right sheet c write(*,*) ' ' c write(*,*) 'Approximate q_n^{k,b}, using the guiding centers' c write(*,*) ' ' do nkp=1,nkpts2 do m=1,nbans do nn=1,nntot(nkp) rnkb(m,nkp,nn)=0.0 brn=0.0 do ind=1,3 brn=brn+bk(ind,nkp,nn)*rguide(ind,m) end do rnkb(m,nkp,nn)=rnkb(m,nkp,nn)+brn end do end do end do do nkp=1,nkpts2 do n=1,nbans do nn=1,nntot(nkp) pherr=dimag(log(csheet(n,nkp,nn)*cm(n,n,nkp,nn))) & -sheet(n,nkp,nn)+rnkb(n,nkp,nn)-dimag(log(cm(n,n,nkp,nn))) if (abs(pherr).gt.3.1416) then write(*,'(3i4,f18.9,3f10.5)') nkp,n,nn,pherr,(bk(i,nkp,nn),i=1,3) end if end do 100 continue end do end do return end From elias.assmann at gmail.com Fri Nov 15 17:47:26 2013 From: elias.assmann at gmail.com (Elias Assmann) Date: Fri, 15 Nov 2013 17:47:26 +0100 Subject: [Wannier] guiding centers In-Reply-To: <52862833.2090503@epfl.ch> References: <5286007F.2050104@gmail.com> <52862833.2090503@epfl.ch> Message-ID: <5286501E.1030308@gmail.com> Dear Nicola, Dear Wannier90 developers, Thank you for your prompt and informative reply. I am taking this off-list for now, but feel free to bring it back to the mailing list, if you prefer. I see from the user guide that the guiding centers are disabled by default. In your experience, how important is this functionality in practical calculations? I have a specific reason for asking about this. I am involved with the wien2wannier interface, and right now I'm preparing it for the next release. I wonder if I should make the effort to enable use of the guiding centers. More flexible initial projections are one of the things I would like to implement for the next release. However, I will still force initial projections to be centered on atoms, since projections centered at arbitrary points are a little more involved with an APW basis. In particular, I am planning to implement rotations of the spherical part of the projections. If I may, I would like to use this opportunity to ask also about the other projection features pw2wannier90 has but wien2wannier does not, i.e. (1) projections centered at arbitrary points, and (2) ?r=? and ?zona=?. How important is it to support these features, in your opinion? I will not implement them for the next release, but maybe we should keep them in mind for the future. Since these are only initial guesses, one might suppose that you can get along fine with ?standard? atom-centered projections. But consider a case like fcc-Cu, which you cite in your review (Fig. 8). Is it necessary to allow interstitial initial guesses to get the interstitial Wannier centers? Elias From xuweiq.e.d at gmail.com Sun Nov 17 13:07:28 2013 From: xuweiq.e.d at gmail.com (Wei Xu) Date: Sun, 17 Nov 2013 20:07:28 +0800 Subject: [Wannier] The relationship between the optical conductivity tensor and the dielectric tensor Message-ID: Dear developers and users: Just a small question. I know that postw90.x can calculate optical conductivity tensors. I'm wondering whether the familiar relationship between optical conductivity tensor and dielectric tensor (which is shown below) still holds here? Since the dielectric tensor for bulk Si which I obtained from the calculated optical conductivity in that way shows considerable off-diagonal elements whereas the macroscopic dielectric matrix for bulk Si is expected to be diagonal. [image: Inline image 1] Of course, an inproper parameter setting could lead to inaccurate results, but for now what's more bothering me is whether the relationship above is still true here. Thank you in advance for any help you could offer :) Wei Xu Ph.D. student Dept. of phys., Fudan University No.220 Han Dan Rd, Shanghai, 200433, China -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: 12121.png Type: image/png Size: 6813 bytes Desc: not available URL: From zhuxg02 at gmail.com Wed Nov 27 04:19:41 2013 From: zhuxg02 at gmail.com (Xiegang Zhu) Date: Wed, 27 Nov 2013 11:19:41 +0800 Subject: [Wannier] memory usage issue during wannierise Message-ID: Dear All, Recently I was using vasp and wannier90 to generate the MLWFs for Bismuth (two atoms in the Rhombohedral lattice). The calculations were done on my laptop (Ubuntu 12.04 64bit) with i7 intel processor and 8G memory. What confused me a lot was the huge memory usage during the convergence calculation during the wannierise process. The memory that wannier90.x used just increased very quickly until my laptop halted. I am wondering whether this is just because I do not have enough memory for my laptop or if there is some funny mechanism that prevents the release of the memory during the calculaton. Here I also post the *.win fill as follows: ---------------------------------------------------------------------------------------- num_wann = 16 num_bands = 30 dis_num_iter = 2000 num_iter = 1000 begin projections Bi: s Bi: p end projections spinors = .true. begin unit_cell_cart 2.6450993 0.0000000 4.0554329 -1.3225496 2.2907232 4.0554329 -1.3225496 -2.2907232 4.0554329 end unit_cell_cart begin atoms_cart Bi 0.0000000 0.0000000 0.0029256 Bi 0.0000000 0.0000000 5.6933354 end atoms_cart mp_grid = 17 17 17 begin kpoints 0.000000000000 0.000000000000 0.000000000000 ...... end kpoints ---------------------------------------------------------------------------------------- Any help or suggestion will be greatly appreciated. Best regards, Xiegang -- Xiegang Zhu Department of Physics and Astronomy Aarhus University Ny Munkegade 120, Building 1520 8000 Aarhus C Denmark Phone: +45 91 19 37 00 E-mail: zhuxg02 at gmail.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.yates at materials.ox.ac.uk Wed Nov 27 12:01:58 2013 From: jonathan.yates at materials.ox.ac.uk (Jonathan Yates) Date: Wed, 27 Nov 2013 11:01:58 +0000 Subject: [Wannier] memory usage issue during wannierise In-Reply-To: References: Message-ID: <78877D222077A642A4CFAB09BD0077E027170EE1@MBX10.ad.oak.ox.ac.uk> On 27 Nov 2013, at 03:19, Xiegang Zhu wrote: > Dear All, > > Recently I was using vasp and wannier90 to generate the MLWFs for Bismuth (two atoms in the Rhombohedral lattice). The calculations were done on my laptop (Ubuntu 12.04 64bit) with i7 intel processor and 8G memory. > > What confused me a lot was the huge memory usage during the convergence calculation during the wannierise process. The memory that wannier90.x used just increased very quickly until my laptop halted. I am wondering whether this is just because I do not have enough memory for my laptop or if there is some funny mechanism that prevents the release of the memory during the calculaton. Xiegang, The memory should stay constant during the wannierise routines, and be comparable to the estimate Wannier90 reports at the start of the calculation. We have had the occasional report of this in the past, and it has always been due to a problem with the compiler or the math libraries. (one of the intel 13.1 compilers had such a problem). Can you check on a different machine / compiler. You might need to upgrade / downgrade the compiler on your laptop. Yours Jonathan -- Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK tel: +44 (0)1865 612797 http://users.ox.ac.uk/~oums0549/ From Krzysztof.Kosmider at inl.int Thu Nov 28 18:44:11 2013 From: Krzysztof.Kosmider at inl.int (Krzysztof Kosmider) Date: Thu, 28 Nov 2013 17:44:11 +0000 Subject: [Wannier] Atomic centres translations and lack of WFs localization Message-ID: Dear all, I calculate with VASP a nanotube (along z axis) made of 64 atoms and then calculate 242 WFs by Bloch states projection over atomic orbitals (p and d) with radially oriented z-axis (x-azis along a nanotube). (I did before successfully monolayer wannierization) If I switch-on guiding centres (guiding_centres=.true.) and avoid minimization procedure (num_iter=0), I obtain a basis of WFs with a spread not greater than 2.0 A^2. However, without guiding centres (guiding_centres=.false.) all atomic centres are translated significantly (so that a sum of centres lies on the x=0.0 y=0.0 axis) and, most importantly, some of the WFs loose localization (spread about 100 A^2). The same happens when I keep guiding centres but allow minimization (even for num_iter=1). With such a not maximally localized basis I obtain meaningful information about electronics of the system. I wonder, however, what is a reason of those translations and why this causes a lack of WF's localization? How to perform total spread minimization in my case? Thank you in advance, Krzysztof Ko?mider INL - International Iberian Nanotechnology Laboratory Av. Mestre Jos? Veiga s/n 4715-330 Braga Portugal email: kkusiu at gmail.com Privileged and Confidential E-mail does not guarantee the security of communications and INL does not accept any liability whatsoever that may arise as a result. This e-mail is intended only for the recipient and contains confidential information, subject to the professional duty of secrecy. Any disclosure, copying or unauthorized use of such information is contrary to law. If you receive this message in error, please notify us immediately and delete the original and any copies and attachments. International Iberian Nanotechnology Laboratory -------------- next part -------------- An HTML attachment was scrubbed... URL: